Ubiquitination‐Related Diagnostic Biomarkers for Diabetic Nephropathy: Insights From Multiomic Analysis, Drug Docking, and Experimental Validation

This study aimed to identify ubiquitination-related diagnostic biomarkers in diabetic nephropathy (DN), explore their molecular mechanisms and candidate therapeutic agents, and further validate their expression and functions through experiments conducted in vivo and in vitro. The GSE30122 dataset from the GEO database was used to screen for differentially expressed genes (DEGs). Ubiquitination-related genes (URGs) were derived from the GSEA database. Weighted gene coexpression network analysis (WGCNA) was applied for the identification of ubiquitination-associated gene modules. Least absolute shrinkage and selection operator (LASSO) logistic regression and support vector machine-recursive feature elimination (SVM-RFE) algorithms were used to screen for ubiquitination-related diagnostic biomarkers in DN. A nomogram diagnostic model was established. ssGSEA was applied to analyze how these biomarkers are related to immune cell types and immune processes. Different ubiquitination-related DN subtypes were identified using consensus cluster analysis. The single-cell dataset GSE195460 was used to explore the cellular distribution of diagnostic biomarkers. Drug prediction and molecular docking were conducted to identify promising therapeutic agents. In vivo (mouse and human DN tissues) and in vitro (NLK-overexpressing podocytes) validation was conducted. Five ubiquitination-related diagnostic biomarkers (TRIM22, RCN1, ZNF204P, LPGAT1, and NLK) were identified, and the nomogram model showed good diagnostic efficiency. Immune-related analysis revealed that these diagnostic biomarkers were significantly associated with immune cells and immune processes. DN samples were clustered into three ubiquitination-related subtypes, with the high-expression subtype exhibiting lower immune infiltration. The molecular docking results indicated that quercetin had favorable binding energies with TRIM22, LPGAT1, and NLK and that resveratrol had good binding energy with RCN1. In vivo validation demonstrated that TRIM22, RCN1, and LPGAT1 were upregulated, whereas NLK was downregulated in the renal tissues of both humans and mice with DN. In vitro experiments confirmed that NLK overexpression ameliorated high glucose-induced podocyte injury and apoptosis by increasing the expression of podocyte markers (podocin and WT1) and Bcl-2 (an antiapoptotic protein) while concurrently reducing the expression of Bax and caspase-1 (proapoptotic proteins). TRIM22, RCN1, ZNF204P, LPGAT1, and NLK are potential diagnostic biomarkers for DN, and the constructed nomogram model has good diagnostic performance. Diagnostic biomarkers are strongly associated with immune cells and immune processes. Quercetin and resveratrol are potential therapeutic agents targeting these biomarkers. NLK may play a protective role in DN by alleviating podocyte injury and apoptosis.